Y. Gil et al. (Eds.): ISWC 2005, LNCS 3729, pp. 872 – 886, 2005.
© Springer-Verlag Berlin Heidelberg 2005
LKMS – A Legal Knowledge Management System
Exploiting Semantic Web Technologies
Luca Gilardoni, Chistian Biasuzzi, Massimo Ferraro, Roberto Fonti,
and Piercarlo Slavazza
Quinary - Via Pietrasanta 14 – 20141 Milan – Italy
Abstract. Semantic Web, using formal languages to represent document
content and providing facilities for aggregating information spread around, can
improve the functionalities provided nowadays by KM tools. This paper
describes a Knowledge Management system, targeted at lawyers, which has
been enhanced using Semantic Web technologies. The system assists lawyers
during their everyday work, and allows them to manage their information and
knowledge. A semantic layer has been added to the system, providing
capabilities that make system usage easier and much more powerful, adding
new and advanced means for create, share and access knowledge.
After years of hype, there is clear evidence of an up-take of knowledge management
in corporations. Today, knowledge is recognized as a strategic resource, with major
key drivers being the need to cut time to market and the fear of missing business
opportunities in a global market where companies have to cope with new products
and services. At the same time, there is a general acknowledgement that existing
technology behind most knowledge management products has somehow reached its
limits. Current knowledge management systems are indeed still mostly built on top of
conventional document management systems, without real 'understanding' layers.
Tools are mostly designed as aids to human centered activities with a set of low level
tools needing human guidance to deliver results . Albeit relying on web
technologies, built as intranet portal tools, current state of the art does not really
leverage the expected potential of the semantic web. To go to the next step, we should
move towards a architecture and an infrastructure providing a foundation for new
generation services, semantically aware tools and proactive agents, able to better
support human actors.
Technology born to support the development of the Semantic Web may be used to
build such foundation. Moreover, the 'inside web', that is the web constituted by
intranets, KM environments, portals, is worth to users as much – and often more –
than the 'web out there'. If we consider the whole world of professional users within
corporation – and to some extent even some virtual community build within closed
spaces – we find out a huge amount of information available. The fact that such
information is not available to the general public is scarcely relevant, as whenever we
consider ourselves, we find that, in our space of accessible information, the outer web
LKMS – A Legal KMS Exploiting Semantic Web Technologies 873
and inside web often play an equally relevant role. What's even more notable is that
the relevance is often in our capacity to connect internal information and external one.
A collaborative environment, such as those currently found behind most intranets,
could provide a natural place to add semantic capabilities, while the organization
work which is behind most intranet initiatives provides the economic support and
impulse to add what's needed – organized information and knowledge – to implement
the semantic layer.
The system described in this paper is centered around this assumption. Based on a
long experience in building advanced knowledge management systems, and derived
from research made in the framework of the Dot.Kom project (), we built an
enhanced solution integrating such a semantic layer into an existent KM environment.
The semantic layer is founded over an ontology repository supporting knowledge
integration and fusion and acting as the common glue for share and reuse services for
knowledge management. Ontologies indeed play a key role in the context of the
Semantic Web: they formalize the knowledge about the concepts related to the
“world” of interest. Once the proper knowledge framework is defined through
ontologies, one can identify in documents instances of the concepts described, and
relations between them. Accordingly, they also play a key role in supporting
knowledge management tools, making them a bit more ‘knowledge aware’. Semantic
annotations, whether manually generated, or derived by information extraction
techniques or other automatic processes, can provide a major framework for
generating, preserving and sharing knowledge. Annotations provide the basis for
advanced information retrieval, and for providing proactive services.
We will describe here a specific vertical solution targeting law firms. The
described system has already been deployed in a major Italian law firm, and is
currently a key component of our company offering for the legal market.
2 Knowledge in the Legal World
Law is a knowledge-based profession. Since law firms and law departments are
knowledge-based organisations, knowledge management becomes critical to their
continuing success. A knowledge management system enables lawyers to work more
efficiently and to provide legal services quicker than ever before. By creating
processes to support and facilitate the identification, capture and dissemination of a
firm’s knowledge, knowledge management systems leverage a law firm collective
The legal industry has faced significant pressures in recent years, making
knowledge management a business imperative. In the age of instant communication,
lawyers have been forced to find quicker ways to deliver traditional legal services.
Law firm clients have become very sophisticated buyers of legal services and
therefore they expect a faster turnaround time.
Several components of knowledge management, such as precedent libraries or
work product repositories, already exist in law firms. Innovative law firms however
are already working to find a more efficient way to work, leveraging the knowledge
of their experts by delegating work to more junior staff and hence looking for better
ways to improve knowledge sharing and exploitation processes.
874 L. Gilardoni et al.
Work of professionals within a law firm – or a legal department in a corporation -
ultimately leads to production of documents: acts, contractsor opinions. In this sense
work processes are document centric. This is one of the reasons why most KM
solutions targeting law professionals focuses on document management issues.
However, from the point of view of knowledge building and sharing, what is really
relevant is the intellectual process carried on to delivery the document.
The outcome of this process is constrained on one side by the task and the specific
matter, from the other by contextual knowledge.
The context should be maintained, because
it is this contextual knowledge that enables, for
example, to maintain and revise documents (..
this clause was made this way because of that
law …; if the law is later amended, or a
different interpretation given by the supreme
court, that clause has to be revised in future
contracts and effects on old contracts has to be
evaluated). Legal documents, moreover, are
inherently interrelated; and so may be the
process that leads to them. A contract – legal act binding two or more parties - may be
designed taking legal opinions into account, and it may be in turn the source of a case
which leads to production of legal acts discussing it, these in turns taking other
opinions into account. Legal opinions - written by lawyers on request of customers
who need advise on some topic - are also based in turn on decisions taken in courts on
specific cases. Being able to keep track of context hence results to be of paramount
The context is given as well by the work process. Lawyers, as many other
professionals, are compelled by the need to share knowledge and competencies.
Findings derived by analysis of a court sentence have to be somehow saved for usage
by other members of the firm. Too many times people end up redoing the same work
as the guy next door in the office. Even a search made against a database looking for
specific cases may be reused in similar cases.
The more a law firm tends to specialize in specific sector, which is often the case,
the more sharing this kind of knowledge gives the competitive advantage. Specific
law firm knowledge is so relevant that the area of practice a law firm is specialized in
highly qualifies the firm. A primary concern and a major activity directly functional
to the primary process is therefore to keep this background knowledge up to date
through knowledge maintenance processes.
One way to share this knowledge is by similarity, which is often the approach
taken by case based reasoning systems, which sometimes work rather reasonably.
Point is that current tools reason by similarity at the textual level, which may work
reasonably to find out similarity in matter (e.g. two employment contracts for
managers with similar bonus plans) but can hardly support in linking at the clause
level (to stay within contracts) or to maintain connections to supporting cases.
Similarity is not taken to the conceptual level, and hidden links and background
knowledge are … just treated as hidden.
Hence knowledge management system to prove effective must support, other than
‘conventional’ search, a way to annotate and hyperlink elements to the surrounding
LKMS – A Legal KMS Exploiting Semantic Web Technologies 875
context, and be able to navigate and search hyperlinks. To be usable, however, the
system must be able to support automatic (at least partially) hyper-linking, and make
easy to manage annotations.
A rich knowledge layer and semantic web technologies provide the foundation to
enhance existing knowledge sharing environments supporting these functionalities. In
a sense, this is not surprising, as the same rationale (adding a semantic layer to
enhance sharing providing a better user experience) is also behind the Semantic Web.
Moreover, in a world where more and more information is going online, and where a
number of public initiatives (e.g. NormeInRete, see ) are strongly driving to make
available public legal information on the web, the more the technology supports
integration of internal law firm material (the 'inside web') with external material, the
more users are likely to take advantage of it.
3 System Description
LKMS (Legal Knowledge Management System) is a collaborative web-based
platform for knowledge management, supporting law firms in managing a document
base and the processes around it. LKMS is a vertical solution for the legal market
built on top of “K@”, a generic KM system developed by Quinary since 2002.
With LKMS users can access and share a common repository of documents while
the system keeps track of people interaction. Documents, including both physical
documents residing inside the law firm, external URLs, notes and Wiki pages, may be
organized according to one or more taxonomies, supporting multiple inheritance
(DAGs): the environment provides a basic framework for sharing information by
matching the way an organization is structuring its processes.
The core system supports browsing and searching using free text queries and
provides a number of tools to track user behavior (who added a document or a node in
a taxonomy, who added classification links between nodes and taxonomies, who
visited nodes or read documents) to facilitate sharing and keeping track of workgroup
Information Extraction Tools
Fig. 1. LKMS Architecture
876 L. Gilardoni et al.
The system also includes a document drafting component, based on XML
technology, enabling to build template documents as aggregate of clauses and
composition logic. Drafts can then be built from templates by specifying constraints
through a user friendly query answer interface.
The core version of LKMS has been enhanced in 2004, now it is able to maintain
the association between documents and semantic annotations with respect to a formal
ontology according to Semantic Web standards. Figure below outlines the overall
Semantic Layer Overview. The core KM framework has been enriched with the
SemantiK plugin to provide a semantic layer over documents repository. SemantiK is
a platform featuring presentation, editing, integration, and searching of knowledge
expressed through the RDF language. SemantiK has been integrated with LKMS as a
plugin, allowing for connecting annotations to documents on evidence that, in most
cases, annotations are motivated by or related to document content.
Storage and inferencing over annotations is given by an underlying RDF repository
(Sesame, see ). The main purpose of SemantiK is to provide a middleware for
high-level access to an RDF knowledge base, supported by a knowledge integration
layer, and a web GUI for maintenance of RDF annotations tailored to end users
habits. Main functionalities include:
ontology supported GUI for viewing, browsing and editing annotations
through web forms;
support for semantic search;
support for knowledge integration and resources disambiguation;
interface and support for automatic annotation extraction from documents.
Fig. 2. SemantiK Internal Architecture
The whole architecture has been built to tackle flexibility in handling of specialized
ontologies, with respect to both presentation issues and semantic integration. LKMS
comes with a precompiled legal ontology, expressed in RDFS1, which may however
1 Enhanced with few custom meta-properties – porting to OWL is being considered.